1 Introduction
Sym-bol | Definition |
---|---|
s
| The source node |
t
| The sink node |
h
| The current node |
Distance from source node to sink node | |
D
h→t
| Distance from current node to the sink node |
T
hop
| The average hop delay at a sensor node |
T
QoS
| The application-specific end-to-end delay objective |
T
h→t
| The reserved time credit for the data delivery from current node to the sink node according to TQoS |
t
s→h
| Data packet's experienced delay up to current node |
H
h→t
| The desired hop count from current node to the sink node according to TQoS |
D
hop
| The desired hop distance for next-hop-selection in MGR |
E
hop
| The energy consumption for one-hop data delivery |
E
ete
| The end-to-end energy consumption for a successful data delivery |
2 Related work
2.1 Wireless multimedia sensor networks
2.2 Geographic routing
2.3 QoS provisioning for delay sensitive traffic in WSNs
3 Architecture of mobile multimedia sensor networks
-
Traditional WSNs have the intrinsic characteristic of scalar data collection (e.g., temperature, humidity, air pressure, etc.), which is hard to elaborate some complicated events and phenomena. In MMSNs, multimedia sensor nodes can provide more comprehensive information such as pictures, text message, audio or videos.
-
The merging of mobility into multimedia sensor nodes further improve the network performance, such as locating mobile nodes to an optimal positions for fast multimedia services, approaching targets for enhanced event description with high-resolution image or video streams, the additional capability for exploring a larger area of sensor nodes to disseminate multimedia streams, as well as various advantages in traditional mobile sensor networks (e.g., load balancing, energy efficiency, improving fairness on the data collection, and coverage optimization, etc.)
-
Though the mobility of multimedia sensor node provides the advantage, the network topology becomes dynamic, which brings difficulties in both the data communication and data management.
4 Illustrative application for MMSNs
5 Mobile multimedia geographic routing
5.1 Analysis of delay-energy trade-offs
5.1.1 Analysis of one-hop delay
-
Queuing delay: For the sake of simplicity, we assume a stable packet rate in our network. Then, queuing delay is considered to be a constant for each hop, which is denoted by T q .
-
Processing delay: With respect to processing delay, we assume that each node incurs similar delay to process and forward one packet with constant length. The processing delay is denoted by T p .
-
Propagation delay: This parameter can be neglected when compared to the other delays.
-
Transmission delay: We assume that the size of a data packet does not change between a source-sink pair, its transmission delay (denoted by Ttx) remains constant between any pair of intermediate sensor nodes.
5.1.2 The end-to-end energy consumption
5.1.3 Energy-delay trade-off
5.2 End-to-end delay objective
5.3 Calculating the desired hop distance at current node
5.4 Strategic location for next-hop-selection
5.5 Next-hop-selection in MGR
begin |
---|
Notation
|
h is the current node to select the next hop node; |
V
h
is the set of node h's neighbors in the forwarding area; |
POS
h
is position of the current node; |
POS
t
is position of the sink node; |
initialization
|
calculate Th→tbased on TQoS and ts→h; |
calculate Hh→tbased on Th→tand Thop; |
calculate D
h
→
t
based on POS
h
and POS
t
; |
calculate Dhop based on Dh→tand Hh→t; |
for each neighbor j in V
h
do |
end for
|
for each neighbor j in V
h
do |
if ΔD
j
= min { ΔD
k
|| k ∈ V
h
} then |
select j as NextHop; |
break; |
end if
|
end for
|
Return j; |
6 Performance evaluation
-
End-to-end Packet Delay: It includes all possible delays during data dissemination, caused by queuing, retransmission due to collision at the MAC, and transmission time.
-
Energy Consumption: the energy consumption for a successful data delivery, which is calculated according to Equation (1).
-
Average Energy Consumption: it is a running mean of ordinate values of input statistic, which is obtained by the statistics collection mode of "Average Filter" in OPNET simulation [28].
-
Lifetime: It's the time when the first node exhausts its energy.